SensiCut: Material-Aware Laser Cutting Using Speckle Sensing and Deep Learning
summary
A compact hardware add-on uses lensless speckle camera imaging and a deep-learning classifier to sense surface microstructure and map sheets to laser power/speed settings, flagging unsafe materials like PVC. The 3D‑printed mount and off‑the‑shelf laser pointer weigh 114 g and attach to the cutter head; the system auto‑samples outlines to produce auto-partitioning of multi‑material objects and reliably selects cutting/engraving parameters across orientations and lighting.
AI-written summary · about
chapters (8)
- 0:00Introduction to SensiCut
- 1:56Problems with current laser cutting
- 3:24Limitations of prior solutions
- 4:56Speckle imaging technique explained
- 6:08Hardware setup and integration
- 6:32User interface and prototyping
- 7:39Engraving multi-material objects
- 8:51Conclusion and future vision
related talks
05mLaser Cutter for STEM and Innovators
17mLaserStacker: Fabricating 3D Objects by Laser Cutting and Welding
17mCatch Me If You Can: Laser Tethering with Highly Mobile Targets
17mSpringFit: Joints and Mounts that Fabricate on Any Laser Cutter
19mAirCode: Unobtrusive Physical Tags for Digital Fabrication
21mAmphiLight: Direct Air Water Communication with Laser Light
20mOff-Line Sensing: Memorizing Interactions in Passive 3D-Printed Objects
15mMakers' Marks: Physical Markup for Designing and Fabricating Functional Objects